See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228018214

Genome Mapping

Chapter · April 2001

DOI: 10.1038/npg.els.0001467

CITATIONS READS
5 8,985

1 author:

Paul H. Dear
Medical Research Council (UK)
121 PUBLICATIONS   5,724 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Developing the next generation of genome editing tools for human use. View project

Neonatal meningitis View project

All content following this page was uploaded by Paul H. Dear on 19 May 2014.

The user has requested enhancement of the downloaded file.

Genome Mapping Introductory article

Paul H Dear, MRC Laboratory of Molecular Biology, Cambridge, UK Article Contents

. Introduction
Genome maps pinpoint the location of specific features on the chromosomes of an . Markers for Genome Mapping
organism. They are essential tools in identifying genes responsible for diseases or traits, for . Methods for Genome Mapping
comparing the genomes of different species, and for complete genome sequencing. . Map Integration

Introduction
A genome map, like any other map, defines the relative
positions of features that are of interest, or which can serve
as reference points for navigation. The features that are
located on a genome map are collectively referred to as
markers, and can include both genes and noncoding
sequences, as outlined below.
Many methods exist for making genome maps, differing
in the types of markers that can be mapped and in their
accuracy. Only one technique (fluorescence in situ hybri-
dization) allows the position of markers in the genome to
be observed directly. In all other methods, a less direct
approach is necessary. A common theme in these indirect
methods is that some means is first used to isolate a portion
of the genome from the remainder (for example, by cloning
in bacterial cells), and to identify the markers present in
that portion. The markers found can then be assumed to be
physically linked in the genome (for example, all markers
found on a single cloned fragment must lie consecutively).
By analysing many such samples of the genome, the order
and spacing of all markers can be inferred.
Good maps are invaluable in many aspects of genome
analysis. Genetic linkage maps (see below), although often
of low resolution, are unique in that they allow the position
of a gene to be determined even when nothing about the
gene is known except for its phenotypic effect. Such maps,
therefore, are often the first step towards isolating the gene
responsible for a disease or for a specific trait. Comparison
of genome maps of different species can reveal conserva-
tion of the order of genes over large regions of chromo-
somes. Such comparative mapping can give insight into
genome evolution, and can provide clues as to the location
of a gene in one species if its position in the genome of
another is known. Higher resolution maps can direct the
early stages of genome sequencing or, if accurate enough,
can help to guide the assembly of the complete sequence
and ensure its long-range continuity and freedom from
gaps.
Some confusion exists between the terms ‘genome
mapping’ and ‘gene mapping’. Although there is some
overlap of meaning, we can define ‘gene mapping’ as the
locating of one or a few genes of particular interest, within
the framework of a more global ‘genome map’. This article
is concerned largely with genome mapping.
Markers for Genome Mapping
Almost any identifiable feature of the genome can serve as
a marker in mapping (Figure 1). The markers most
commonly used are simply short, unique fragments of
known deoxyribonucleic acid (DNA) sequence, which can

STS µSat. RFLP Contig Probe Locus

CACCGCTAGCCGACTCAGTG CGTCTAGCCTAGCCTACTAG

CACCGCTAGCCGACACAGTG ACACACACACACACAC CGTCTAGCCTAGCCTACACG

Figure 1 Markers used in genome mapping. This hypothetical composite map shows several types of marker located on the genome (heavy horizontal
line). STS: sequence-tagged site, a region of known sequence which can be amplified and detected by the polymerase chain reaction using flanking
sequence-specific primers (arrows); if the STS lies within a gene, it is an ‘expressed sequence tag’ or EST. mSat.: microsatellite, an STS that encompasses a
simple, tandemly repeated sequence; the number of tandem repeats can vary from one individual to the next, making the marker polymorphic. The length
of the PCR product produced using the flanking primers (arrows) will reflect the number of repeats. RFLP: a polymorphic restriction site, present in some
individuals but absent in others. If genomic DNA is digested with the appropriate restriction enzyme, the sizes of the fragments will reflect the presence or
absence of the site. Contig: a series of cloned DNA fragments that overlap one another in a contiguous series. The absolute position of the contig within the
genome is known only if one or more of the clones contains a marker mapped by independent means. Probe: a labelled, cloned DNA fragment whose
position in the genome can be observed directly by hybridizing it to metaphase chromosomes and observing it under a microscope. Locus: a gene or other
sequence known only by its phenotypic effect. Its location in the genome can be inferred from the pattern of inheritance of the phenotype.

ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net 1
 Genome Mapping
be amplified and detected in a DNA sample using the
polymerase chain reaction (PCR). Such sequence frag-
ments are collectively called ‘sequence tagged sites’ or
STSs. Often, the STS will be some arbitrarily chosen piece
of DNA (obtained by sequencing a clone picked at random
from a genomic clone library), and has no particular
biological significance. However, an STS can originate
from a coding sequence (for example, by sequencing a
clone from a complementary DNA (cDNA) library), in
which case it is an ‘expressed sequence tag’ or EST. A third
category of STS consists of a tandem repetition of a simple,
short motif (for example, CACACACA...) flanked by a
unique sequence. Such ‘microsatellites’ can be detected by
PCR using their unique flanking sequence but often display
length polymorphism from one individual to the next, as
the number of repeat elements varies. Such polymorphism
is of key importance in genetic linkage mapping.
Cloned DNA fragments can be mapped, and therefore
act as markers, even when little or nothing is known of their
sequence. Techniques for mapping such cloned fragments
include physical mapping and fluorescence in situ hybridi-
zation. If the anonymous cloned fragment contains a
restriction site whose location differs from one allele to the
next (a restriction fragment length polymorphism or
RFLP), then genetic linkage mapping can be used.
Finally, phenotypic variations act as genetic linkage
markers even when nothing whatever is known of the
underlying DNA sequence. In these cases, it is the
phenotypic trait that is the observable marker, from which
the presence of the allele at the DNA level is inferred.
Methods for Genome Mapping
Genetic linkage mapping
In organisms that reproduce sexually, meiosis breaks the
parental chromosomes at random, recombines the frag-
ments and segregates the shuffled chromosomes into
gametes, and thence to offspring. If two markers, A and
B, lie close together on a chromosome, then it is unlikely
that a meiotic break will occur between them. Hence, A and
B will seldom recombine, and will usually segregate
together (cosegregate) to the same gametes and hence to
the same offspring. In contrast, if markers B and Z lie at
opposite ends of a chromosome, it is much more likely that
a meiotic break will fall between them, whereupon they
may segregate independently. Markers lying on different
chromosomes will also segregate independently. Hence,
the distance between any two markers is reflected by their
recombination frequency: closely linked markers recom-
bine rarely (or cosegregate often) while distant markers
recombine often, and hence cosegregate no more than
expected by chance – 50% of the time. This provides a way
2 ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net
to estimate the distances between pairs of markers and, if
all pairwise distances are known, a map can be made.
There is one complicating factor, however. Meiotic
recombination occurs by an exchange of homologous
segments between chromosomes. Hence, some way must
be found to distinguish between the copies of the markers
on each parental chromosome, so that the pattern of
segregation can be followed. This is only possible in the
case of polymorphic markers, and recombination can be
detected only in those cases where the parent is hetero-
zygous at the relevant loci. In addition, three generations
(grandparents, parents and offspring) are normally re-
quired to be able to deduce completely the pattern of
segregation (Figure 2).
Many types of polymorphic markers exist. In the earliest
genetic linkage maps, the markers were not defined DNA
sequences but variable traits which were assumed to reflect
underlying (and then undetectable) polymorphisms. Such
‘trait mapping’ is still performed, and has the great
advantage that it can identify the relative positions of
genes based solely on their phenotypic effect. This is
exploited, for example, in mapping disease loci as a first
step toward identifying the gene responsible.
More often, though, genetic linkage maps are made
using markers characterized at the sequence level. The
earliest such markers were restriction fragment length
polymorphisms or RFLPs – regions of sequence in which
the alleles differ in the distribution of sites for a particular
restriction enzyme. RFLPs can be detected by digesting
genomic DNA with the restriction enzyme, resolving the
fragments electrophoretically, Southern blotting and
probing with a labelled piece of DNA complementary to
the region of interest. Different alleles then appear as bands
of different sizes on the autoradiograph.
RFLPs have largely been superseded in linkage mapping
by microsatellites, regions of simple repeated sequence
(such as CACACACA...) in which the number of repeats
differs between alleles. Alleles are scored by measuring (on
gels) the length of a PCR product produced by primers
which anneal to the unique sequence on either side of the
microsatellite.
Sequence polymorphisms of a single nucleotide can be
detected by PCR- or hybridization-based methods. Such
single nucleotide polymorphisms or SNPs are extremely
common in the human genome, occurring once every few
hundred bases on average. However, SNPs are more often
identified once a region of the genome has been mapped
and sequenced, rather than serving as markers for the
initial genetic linkage mapping.
The analysis of linkage data to produce maps is far from
simple (especially where many markers are to be mapped)
and several mathematical methods exist. The first step is
normally to estimate the distances between all possible
pairs of markers in the study. In the ideal case, this can be
done by finding the proportion of meiotic events in which
recombination occurs between the two markers. However,
 Genome Mapping

AB Z AB Z AB Z AB Z

ab z ab z ab z ab z

AB Z AB Z AB Z AB Z

ab z ab z ab z ab z

AB z AB z Ab z AB z

+ + + +

ab Z ab Z aB Z ab Z

Figure 2 Principle of genetic linkage mapping (only one chromosome is shown). Diploid parental cells (top row) carry polymorphic marker alleles A, B
and Z on one chromosome and alleles a, b and z on the other homologue. Meiotic recombination can occur at any point (indicated by the hourglass shape)
along the chromosomes, causing an exchange of markers on either side of the site of crossover (second row). The recombined chromosomes are then
segregated to gametes (bottom). Recombination is unlikely to occur in the small interval between closely-linked markers such as A and B – in this example,
only one of the four parental cells (two of eight gametes) show recombined genotypes Ab or aB. Recombination between widely spaced markers B and Z is
more common, producing many gametes with bZ or Bz genotypes. By genotyping the members of a pedigree, the pattern of recombination can be
deduced, allowing the distances between markers to be estimated.

linkage studies seldom provide such complete data, for
example because not all members of the pedigree are
available for genotyping or heterozygous at the appro-
priate loci. A more approximate method must therefore be
used, in which various possible intermarker distances are
modelled, to find the distance that is most likely to give rise
to the observed pattern of segregation. Once the distances
between all possible pairs of markers have been estimated
in this way, the arrangement of all the markers is found that
best agrees with the complete set of pairwise distance
estimates. Because the number of possible marker orders
increases as the factorial of the number of markers (for
example, 3  2  1 5 6 possible orders for three markers,
but 4  3  2  1 5 24 for four markers, and so on),
rigorous analysis is seldom possible, and some degree of
compromise or approximation is usually required.
Linkage data are most powerful when all markers have
been analysed in a common set of individuals. Only in this
way can all of the pairwise distances between markers be
estimated. For example, if markers A, B and C are mapped
on one pedigree, and markers C, D and E on another, then
the distances A–D, A–E, B–D and B–E cannot be directly
inferred. The CEPH panel (Centre d’Etudes du Polymor-
phisme Humain) is a collection of DNA samples from large
multigeneration families, available to researchers world-
wide. By typing polymorphic markers on this common
resource, the resulting linkage data can be integrated
effectively. Equivalent shared resources have been estab-
lished for many other species.
Genetic linkage mapping is a powerful method, but
suffers from some limitations. In particular, only poly-
morphic markers can be mapped, and the resolution that
can be achieved is limited. In humans, recombination
occurs only every 100 Mb or so on average, limiting the
resolution to around 1 Mb if a few hundred individuals are
genotyped. Finally, it has been found that certain small
regions of genomes (including the human) are more prone
to meiotic recombination than others. Markers situated on
either side of such ‘recombination hotspots’ will recombine
more often than would otherwise be the case, leading to an
overestimation of the distance between them.
Radiation hybrid mapping
In the mid-1970s, Stephen Goss and Henry Harris
discovered that, if cultured human cells were subjected
to high dose of radiation (enough to fragment their

ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net 3
 Genome Mapping
chromosomes) and then fused to unirradiated hamster
cells, hybrids would be produced. Such ‘radiation hybrids’
would initially contain many fragments of human chromo-
somes (in addition to the complete hamster genome), and
hence would express many human proteins. Over time,
however, the unstable human fragments would be lost (and
with them, the human proteins) until only a few were stably
retained. If, in many such hybrids, two human proteins
were often retained together, it was likely that their genes
lay on the same human chromosome fragment. In this way,
the arrangements of genes on the human chromosomes
could be deduced.
This method was little exploited until the late 1980s when
it was updated using modern molecular genetic techniques
by researchers including David Cox, Mike Walter and
Peter Goodfellow. In its current form, hybrids are formed
in essentially the same way between irradiated donor cells
and unirradiated host cells. Hybrids containing donor
chromosome fragments are then identified using selectable
markers (for example, drug resistance genes) from each
species. Many such hybrids are isolated and cultured, to
give a mapping panel of perhaps a hundred distinct hybrid
cell lines, each containing a different set of donor
fragments.
DNA is prepared from each cell line in the panel, and is
screened to determine which donor-derived genome
markers it contains. The markers are usually STSs
(including ESTs), which can easily be detected using
PCR. In this way a table of results is produced showing
which markers are present in each member of the mapping
panel. As in Goss and Harris’ original method, two
markers that are frequently found together in the same
hybrid (i.e. which cosegregate often) can be inferred to lie
close together in the donor genome. In this way, the order
and distance of markers in the genome can be inferred.
Radiation hybrid (or RH) mapping is closely analogous
to genetic linkage mapping. Instead of recombination, it is
the radiation that randomly breaks the chromosomes;
hybrid cell formation replaces meiotic segregation as a way
to shuffle and segregate these fragments. As with linkage
mapping, cosegregation frequencies reflect the proximity
of markers to one another. Therefore, the analytical
techniques used to construct the maps are fundamentally
similar and, with minor modifications, interchangeable.
RH mapping differs from linkage mapping in several
important ways, however. Because the donor fragments
segregate directly into the hybrids (rather than being
exchanged between homologous chromosomes as in
linkage mapping), the method is not restricted to
polymorphic markers. This means that almost any STS
can be mapped. Moreover, the resolution afforded by RH
mapping is far higher than that attainable by linkage
mapping. This resolution depends only upon how finely the
donor chromosomes are broken at the outset, which in turn
can be controlled by varying the dose of radiation. If the
dose is low, the donor fragments are large and therefore the
4 ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net
resolution is poor. A higher dose will break the chromo-
somes into smaller fragments, allowing the pattern of
segregation to be observed on a far smaller scale and
permitting finer maps to be made. Resolutions down to a
few tens of kilobases can be achieved in this way.
RH mapping, nonetheless, suffers from a few limita-
tions. First, the production and growth of suitable hybrids
is far from easy, and has not yet been possible for a wide
variety of species. Second, reliable RH mapping requires
that the radiation-induced breaks and the segregation of
donor chromosome fragments amongst the hybrids should
be uniform for all parts of the donor genome. In fact, this is
known not to be the case: for example, donor fragments
that contain centromeres are more likely to be stably
retained in the hybrid cells than those that do not. Donor
telomeres and certain genes can also bias the retention of
fragments. These biases can lead to distortions and errors
in the resulting maps. Third, problems can arise in mapping
markers that are closely homologous between the host and
donor genomes (as is often the case with ESTs derived from
conserved genes). In such cases, it may be difficult to
distinguish the donor marker in the hybrid from the host’s
counterpart, complicating the analysis.
Despite these shortcomings, however, RH mapping is an
invaluable tool in human and in many other vertebrate
species.
HAPPY mapping
RH mapping overcomes many of the limitations of genetic
linkage mapping. However, its few remaining problems
(difficulties in making hybrids, distortions arising from the
effects of donor centromeres and other sequences, and
difficulty in distinguishing some donor markers from host
sequences) all arise from the in vivo step – the use of hybrid
cells to propagate the donor fragments. HAPPY mapping
overcomes these limitations, being an entirely in vitro
technique. Indeed, it can be thought of as ‘radiation hybrid
mapping without radiation hybrids’.
In HAPPY mapping, DNA is first prepared from the
species whose genome is to be mapped. To prevent
unwanted mechanical breakage to the naked, fragile
DNA molecules, this is normally done by embedding the
living cells in agarose gel, and then treating them with a
solution of detergents and proteases. These diffuse into the
agarose, lysing the cells and stripping away proteins and
other cellular debris, which diffuse out of the agarose; the
long molecules of chromosomal DNA remain trapped and
protected within the agarose.
The DNA is now broken randomly, either by melting the
agarose and mechanically shearing the DNA solution, or
by treating the embedded DNA with gamma radiation
(much as in the initial stages of RH mapping). The result is
a pool of random DNA fragments, whose average size

depends upon the violence of the mechanical shearing or
on the dose of radiation used.
This broken DNA is then diluted to a very low
concentration, and about one hundred samples are
dispensed into separate tubes. The complete set of samples
is referred to as a ‘mapping panel’. Most importantly, each
sample is exceedingly small, containing less than one
genome’s worth of DNA fragments. For example, a
mammalian genome contains about 3 pg (3  10 2 12 g) of
DNA per haploid copy; in this case, each sample would
contain only 1–2 pg of DNA fragments.
These minuscule samples are screened by PCR, to
determine which markers (STSs or ESTs) are present in
each one. (The fact that roughly haploid samples of DNA
are screened using the polymerase chain reaction gives the
method its name.) Because the samples are so small, each
one will contain only a randomly sampled subset of the
markers, rather than the complete genome. Hence, any
particular marker will be present in some, but not all of the
members of the mapping panel. If two markers are close
together in the genome (compared to the average size of the
random fragments), then they will seldom be broken apart
and hence will tend to be found together in the same
members of the mapping panel – they cosegregate. As the
distance between markers increases, it is more likely that
the random breakage will separate them, so that they lie on
separate DNA fragments and are hence less likely to
cosegregate. If the markers are very far apart, they will
always be broken apart during the random breakage, and
hence will show no particular tendency to cosegregate.
As in RH mapping, statistical analysis of the cosegrega-
tion frequencies between markers allows the distance
between them to be estimated. Once all the pairwise
distances are known, the order and spacing of the markers
in the genome can be deduced. The analytical procedures
are effectively identical to those used in RH mapping, and
the same software packages can be used to analyse either
type of data.
The resolution that can be attained depends only upon
how finely the DNA is broken at the outset. Since naked
DNA can be broken into fragments as small as a few
hundred base pairs, the resolution of this method can be
very high, and maps with resolutions of a few kilobases
have been produced.
The in vitro nature of HAPPY mapping eliminates some
of the potential limitations of RH maps. Since no cell
culture or hybrid formation is required, the method can be
applied to any genome. The segregation of DNA fragments
is done by simple dilution, and is therefore not affected by
centromeres, telomeres or other biologically active se-
quences. Since there is no ‘host genome’, the samples can be
screened for markers without interference from homo-
logous host sequences. As a result, HAPPY maps are
relatively immune to artefacts.
The in vitro nature of HAPPY mapping, however, also
imposes certain limitations of its own. The first is that it is
ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net
Genome Mapping
difficult to prepare DNA fragments of more than a few
megabases in size. This in turn means that intermarker
distances of more than a megabase or so cannot be
measured, making it unsuitable for producing very low-
resolution maps in which the markers are widely spaced.
(In contrast, most other techniques are restricted in the
maximum resolution they can attain.) The second limita-
tion is that the DNA samples in the mapping panel are
exceedingly small, yet must be screened (by PCR) for all of
the markers to be mapped. In the analogous stage of RH
mapping, the hybrid cells are cultured to produce large
quantities of DNA before marker screening begins. In
HAPPY mapping, the DNA samples cannot be ‘grown’ in
the same way. Instead, in vitro techniques must be used to
‘preamplify’ the minute samples, replicating many-fold
any marker sequences which they contain. Only after this
preamplification can the samples be treated as a bulk
resource, and screened in turn for many markers. Several
techniques are available for the preamplification step,
including PEP (primer-extension preamplification) and
ligation-mediated whole-genome PCR. This is the most
critical step in HAPPY mapping, as the loss of any markers
during preamplification will make them unmappable.
Nevertheless, HAPPY mapping has proved effective in
making high-resolution maps in a wide variety of species.
Physical mapping
Physical mapping involves finding a contiguous series (or
‘contig’) of cloned DNA fragments which contain over-
lapping portions of the genome. The overlaps define the
positions of the clones relative to one another. If at least
some of the clones contain markers that have been
independently mapped by other means, then the position
of the entire contig in the genome is also known.
The starting point for physical mapping is a library of
cloned genomic fragments, normally prepared by either
random mechanical breakage or partial restriction diges-
tion of genomic DNA. (Complete restriction digestion of
the DNA would mean that no clones could be found that
overlapped across the restriction sites for that enzyme.)
The fragments are usually cloned in bacterial hosts
(normally Escherichia coli), using bacteriophage, cosmid,
plasmid or other vector systems. For physical mapping of
large genomes, it is desirable to use clones containing large
inserts, such as P1 artificial chromosomes (PACs) or
bacterial artificial chromosomes (BACs) which can carry
inserts in excess of 100 kb. Even larger fragments (over
1 Mb) can be cloned in yeast (Saccharomyces cerevisiae)
using yeast artificial chromosome (YAC) vectors; how-
ever, such clones are often unstable, undergoing deletions
or internal rearrangements.
Once the library is established, overlapping clones can be
identified using several different approaches. In STS
content mapping, the library is screened to identify all
5
 Genome Mapping
clones that contain a specific STS marker. Screening can be
done either by PCR using the appropriate primers (in
which case only clones containing the STS will give a PCR
product); or by arraying the clones on a nylon membrane,
lysing them to expose their DNA, and probing the
membrane with labelled DNA corresponding to the STS
in question. In either case, all clones containing a given STS
marker are identified. If two clones are found to contain the
same STS, then they must represent overlapping parts of
the genome (with the STS lying in the region of the
overlap). If one clone contains two different STSs, then
those two STSs must be consecutive in the genome. If
enough clones are characterized for enough STSs, then a
contiguous overlapping path of clones is defined.
A second common approach to physical mapping is by
the use of restriction fingerprinting. In this case, each
member of the clone library is digested in turn with one or
more restriction enzymes, and the sizes of the resulting
fragments are measured by gel electrophoresis. If two
clones are found that have several fragment sizes in
common, then they must represent overlapping parts of
the genome, with the shared fragments coming from the
region of overlap. Again, if enough clones are fingerprinted
then the hope is to find a contiguous series of overlapping
clones.
Whatever the means used to make them, a variety of
problems and artefacts can afflict clone contigs. Many
clones must be screened to find ones that cover as much of
the genome as possible. As more clones are screened, the
number of contigs at first increases (since most new clones
do not overlap with earlier ones, and hence form
independent contigs), and then declines as more clones
are found that overlap and link the earlier ones. Typically,
a coverage of between 5- and 10-fold is required, meaning
that the total DNA content of the screened clones must be
5- to 10-fold greater than the genome to be covered. (For
example, if a genome of 100 Mb is cloned in 100 kb
fragments, then 5000 clones would give 5-fold coverage.)
Even then, there are likely to remain some gaps in the
contig, corresponding to regions of the genome that resist
cloning in the particular host/vector system used. Such
gaps can often be filled by using clones from a different
library (made using a different host/vector system). A
process known as chromosome walking is often used to
help close gaps: sequences are obtained from the ends of
clones flanking the gap, and are used as probes to
specifically identify new clones that extend into or across
the gap.
Physical mapping is also vulnerable to a variety of clone
artefacts. Chimaeric clones (that is, clones in which two
unrelated pieces of DNA from different parts of the
genome have become ligated together) can cause the contig
to ‘jump’ from one part of the genome to another. Clones
that have suffered internal deletions or rearrangements can
also cause errors. Finally, sequences that are duplicated in
the genome can disrupt contig assembly, since two clones
6 ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net
that contain different copies of the repeated motif may
wrongly be assumed to overlap.
Despite its limitations, however, physical mapping has
the unique advantage that, as well as producing a map
giving the relative locations of the clones (and of any STS
markers they are known to contain), it also produces a
tangible resource – the clones themselves – representing
known segments of the genome. These mapped clones can
be returned to for more detailed analysis, or can serve as the
substrate for genome sequencing.
Fluorescence in situ hybridization
All of the preceding methods of genome mapping are
indirect: the order of markers in the genome must be
inferred by analysing the marker content of meiotic
products, hybrid cells, DNA samples or cloned DNA
fragments. In contrast, fluorescence in situ hybridization
(FISH) is an elegantly direct way of observing the physical
arrangement of markers along the chromosomes. Fluor-
escently labelled probes are hybridized to metaphase
chromosomes on a glass slide, and their position observed
using a fluorescence microscope.
The first requirement for conventional FISH is a
population of cells in which a high proportion are in
metaphase, the stage in the cell cycle where the chromo-
somes are condensed as distinct bodies. This is typically
achieved by culturing lymphocytes or fibroblasts in the
presence of a mitogen such as phytohaemagglutinin, then
treating them with colcemid to block the cell cycle at
metaphase. Treatment with bromodeoxyuridine (which is
incorporated into the replicating chromosomes) and
ethidium bromide (which intercalates into the DNA)
improves the morphology of the chromosomes for FISH.
The cells are swollen in a hypotonic salt solution and
stabilized with a fixative. Drops of the cell suspension are
placed on glass slides, to which the chromosomes adhere.
Their DNA is denatured, and is then ready to be hybridized
to the probe.
Probes are normally cloned fragments of genomic DNA
or cDNA, which have been labelled (for example, using
nick-translation) by the incorporation of a hapten such as
biotin or digoxigenin. They are denatured, allowed to
hybridize to the corresponding sequences of the metaphase
chromosomes, and the surplus washed away. Nonspecific
hybridization is normally suppressed by using an un-
labelled competitor such as total genomic DNA. To detect
the probes, the slide is incubated with fluorescently labelled
proteins (such as avidin or anti-digoxigenin antibodies)
which bind to the hapten in the probe and render it visible
in the fluorescence microscope. At the same time, the
chromosomes themselves can be visualized by a fluorescent
counterstain, such as DAPI (4’,6’-diamidino-2-phenylin-
dole), of a different colour to the fluorophore used to detect
the probe.

Figure 3 Dual-colour fluorescence in situ hybridization. The image shows
a spread of male canine metaphase chromosomes, probed with two
plasmid clones which hybridize to the X-chromosome (towards the right of
image). One clone has been visualized using a red fluorophore (Texas red),
the other with green (FITC). Chromosomes have been counterstained with
DAPI to reveal their characteristic banding pattern. Each probe produces
two adjacent spots, corresponding to the two sister chromatids. The
distance between the probes is approximately 20 Mb. Picture courtesy of
Dr H. F. Spriggs.
The resulting images clearly reveal the location of the
bound probe, typically as two adjacent spots correspond-
ing to the two chromatids (Figure 3). The characteristic
banding pattern of the chromosomes gives the absolute
location of the probe in the genome. By using different
combinations of haptens and fluorophores, two or more
probes may be hybridized and detected simultaneously.
Such multicolour FISH allows the order and spacing of the
probes to be determined more accurately than is possible
with independent, single-probe experiments.
Metaphase chromosomes are relatively condensed
(typically containing about 10 Mb of DNA per micrometre
of their length), limiting the resolution of FISH to about
1 Mb. However, hybridization of probes to interphase
nuclei (in which the chromatin is less condensed) provides
considerably higher resolution, although there is no longer
a visible banding pattern to provide an absolute chromo-
somal location. Chromatin fibres released from interphase
nuclei are more extended still, and hybridization to these
(known as ‘fibre FISH’) can allow the relative positioning
of probes to within a few kilobases.
ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net
View publication stats
Genome Mapping
Map Integration
Each of the above approaches to genome mapping has its
own strengths and weaknesses, and the same genome will
often be mapped by several groups using different
techniques. One of the greatest problems in genome
analysis, therefore, is the integration of these disparate
maps to produce a unified picture of the genome.
Alignment of different maps of the same region is
hampered in the first instance because different markers
will usually have been placed on each map. Even where
there are some markers in common, placement of the
remaining, unshared markers can be problematic. For
example, one map may show markers A-X-B in that order,
while another shows markers A-Y-B. Should the inte-
grated map be A-X-Y-B or A-Y-X-B?
Even when most of the markers are common to two
maps, problems remain because the accuracy of different
maps varies, as does the nature of the errors likely to be
present. For this reason, it is common to find that maps
disagree over the local (or sometimes not so local) order
and spacing of those markers that they share. Resolution of
these conflicts requires a detailed consideration of the
strengths and weaknesses of each type of map.
The problem of effective map integration has not been
fully addressed, and researchers interested in a particular
region of the genome will often have to use their best
judgement to make sense of the conflicting maps of that
area. Nevertheless, attempts have been made to present
and unify maps as far as possible; a notable example in the
human genome is the Genome Database (see Further
Reading).
Further Reading
Coulson A and Sulston J (1988) Genome mapping by restriction
fingerprinting. In: Davies KE (ed.) Genome Analysis – A Practical
Approach, pp. 19–39. Oxford: IRL Press.
Dear PH (ed.) (1997) Genome Mapping – A Practical Approach. Oxford:
IRL Press.
Dear PH and Cook PR (1993) HAPPY mapping: linkage mapping using
a physical analogue of meiosis. Nucleic Acids Research 21: 13–20.
Genome Database. [http://www.gdb.org/]
McCarthy LC (1996) Whole genome radiation hybrid mapping. Trends
in Genetics 12: 491–493.
Ott J (1986) A short guide to linkage analysis. In: Davies KE (ed.) Human
Genetic Disease – A Practical Approach, pp. 19–32. Oxford: IRL Press.
Strachan T and Read AP (1999) Human Molecular Genetics, 2nd edn, pp.
244–250. Oxford: BIOS Scientific.
7